Closed fjebaker closed 9 months ago
Two tests failed:
smoke-utility: Test Failed at /home/runner/work/Gradus.jl/Gradus.jl/test/smoke-tests/cunningham-transfer-functions.jl:37
Expression: ≈(measure_ctf(test_ctf1(0.998, 30, 800.0)), 0.17351504941705492, atol = 1.0e-5)
Evaluated: 0.1682889574818813 ≈ 0.17351504941705492 (atol=1.0e-5)
smoke-utility: Test Failed at /home/runner/work/Gradus.jl/Gradus.jl/test/smoke-tests/cunningham-transfer-functions.jl:38
Expression: ≈(measure_ctf(test_ctf1(0.998, 30, 1000.0)), 0.19085561680308122, atol = 1.0e-5)
Evaluated: 0.21563413017178515 ≈ 0.19085561680308122 (atol=1.0e-5)
Those are fairly significant differences. They are also the two really large radii tests, so I think I can track this down to being related to the gtol
function needing some more tuning.
Patch coverage: 91.37%
and project coverage change: -0.48%
:warning:
Comparison is base (
ad4db76
) 68.98% compared to head (6d3a239
) 68.51%.
:exclamation: Your organization needs to install the Codecov GitHub app to enable full functionality.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
There are a number of things I am still not really happy about. I'll document them in an issue so that this PR can be merged.
The thick disc transfer functions seem to integrate mostly okay for default parameters with this PR. The main changes were: